Data-driven decision support for rail traffic control: A predictive approach

Advanced decision support for rail traffic control is significant enhancing the safety and quality of railway transport service. Data-driven methods have shown powerful learning ability wide extensibility prediction, classification, decision-making problems. In this paper, we propose a hybrid prediction model based on deep forest (DF) ensemble to analyze two common actions, i.e., changing dwelling times running times. This basic concept mimic controllers, providing them with advanced decisions/control actions using data-driven algorithms. According approach process split into stages, type action (ToA) number changes (NoC) in (i.e., how many or been changed compared planned ones). The first stage classification problem; thus, DF classifier synthetic minority oversampling technique (SMOTE) employed deal imbalanced data. second stage, regressor treats NoC as numerical variables utilizes information from one, results model, make predictions. proposed calibrated train operation data high-speed lines China. experimental comparative analyses show that method provides timely controllers. These characteristics are imperative supporting dynamic controllers manage traffic.